Proc. Natl. Acad. Sci. USAVol. 93, pp. 9594-9599, September 1996Cell Biology

Altered surfactant function and structure in SP-A genetargeted miceTHOMAS R. KORFHAGEN*t, MICHAEL D. BRUNO*, GARY F. Ross*, KAREN M. HUELSMAN*, MACHIKO IKEGAMIt,ALAN H. JOBEt, SUSAN E. WERT*, BARRY R. STRIPP*§, RANDAL E. MORRIS*, STEPHAN W. GLASSER*,CINDY J. BACHURSKI*, HARRIET S. IWAMOTO*, AND JEFFREY A. WHITSETT**Children's Hospital Medical Center, Department of Pediatrics, Cincinnati, OH 45229-3039; and tDepartment of Pediatrics, Harbor-University of California,Los Angeles, Medical Center, Torrance, CA 90509

Communicated by John A. Clements, University of California, San Francisco, CA, May 10, 1996 (received for for review January 4, 1996)

ABSTRACT The surfactant protein A (SP-A) gene wasdisrupted by homologous recombination in embryonic stemcells that were used to generate homozygous SP-A-deficientmice. SP-A mRNA and protein were not detectable in the lungsof SP-A(-/-) mice, and perinatal survival of SP-A(-/-)mice was not altered compared with wild-type mice. Lungmorphology, surfactant proteins B-D, lung tissue, alveolarphospholipid pool sizes and composition, and lung compliancein SP-A(-/-) mice were unaltered. At the highest concen-tration tested, surfactant from SP-A(-/-) mice produced thesame surface tension as (+/+) mice. At lower concentrations,minimum surface tensions were higher for SP-A(-/-) mice.At the ultrastructural level, type II cell morphology was thesame in SP-A(+/+) and (-/-) mice. While alveolar phos-pholipid pool sizes were unperturbed, tubular myelin figureswere decreased in the lungs of SP-A(-/-) mice. A nullmutation of the murine SP-A gene interferes with the forma-tion of tubular myelin without detectably altering postnatalsurvival or pulmonary function.

Surfactant protein A (SP-A) is an abundant, phospholipid-associated protein in pulmonary surfactant. Murine SP-A isencoded by a single gene located on chromosome 14, at a sitesyntenic with the SP-A locus on chromosome 10 in humans (1).SP-A is synthesized and secreted primarily by type II andbronchiolar cells in the respiratory epithelium. In the alveolus,SP-A forms large oligomers and is closely associated withtubular myelin, the major extracellular form of surfactant (fora review, see ref. 2). SP-A contains a 10-kDa collagen-likeamino-terminal domain and a globular carboxyl-terminal do-main with structural homology to SP-D, mannose bindingprotein, Clq, and other members of the collectin family ofmammalian lectins (3). Various functions related to surfactanthave been ascribed to SP-A, based primarily on in vitrofindings, which include enhancing phospholipid uptake (4),inhibiting surfactant secretion by isolated type II epithelialcells (5, 6), contributing to tubular myelin formation (7),enhancing surfactant spreading, stabilizing phospholipid mix-tures (8) and conferring resistance to protein mediated inac-tivation of surfactant (9). Evidence has also accumulated thatsupports a role for SP-A in host defense. SP-A binds andincreases the uptake of a variety of bacterial and viral patho-gens by alveolar macrophages, activates alveolar macrophagemetabolism, and enhances the uptake of bacterial pathogensby both macrophages and monocytes in vitro (10-13). In spiteof the extensive in vitro evidence that SP-A is involved insurfactant homeostasis and function, there is little directevidence that SP-A plays a critical role in surfactant functionin vivo. To test the role of SP-A in vivo, the mouse SP-A gene

was disrupted by homologous recombination in embryonicstem cells to produce mice lacking SP-A.

MATERIALS AND METHODSGeneration of Null Mutant Mice. A 129J genomic library

packaged in A DASH II (Promega), kindly provided by MarciaShull (University of Cincinnati), was screened to isolate themurine SP-A gene locus. Partial restriction endonuclease mapsand partial sequence analysis of the 129J mouse genomic SP-Agene were highly homologous to the previously published genefrom strain DBA2J (14). A 3.6-kb HindIlI fragment of themurine SP-A gene was cloned into Puc18. The murine SP-Agene fragment digested with NcoI at positions 795 and 1049,and the HindIII/XhoI fragment containing pGKneoBPA wereconverted to blunt ends with Klenow fragment. pGKneoBPAwas cloned between these positions to replace portions ofexons 3 and 4 and intron 3. Orientation of pGKneoBPA wasconfirmed by PstI restriction endonuclease digestion. Herpessimplex virus-thymidine kinase (HSV-TK) was cloned into thePuc18 NdeI site 3' to pGKneoBPA insert. The targeted SP-Agene was prepared as a SalI fragment for electroporation intoembryonic stem (ES) cells.ES cells (clone E14.1) generated from 129 ola mice were

grown on neomycin-resistant mouse embryonic fibroblastfeeder layers. Using '20 jig of the SP-A targeting construct,about 2 x 107 of the cells were electroporated. Targeted EScells were selected by culture in G418 (150 ,ug/ml) andgancyclovir (2,tM). Targeted cells were identified by digestionof genomic DNA with EcoRI, SspI, or BamHI and analyzedseparately with two probes outside of the insert (probe 1 frompositions 4028-4618 and probe 2 from positions 2215-3481)and a third internal probe (positions 769-2297) (see Fig. 1).Clone 87, a rapidly proliferative, undifferentiated ES cellclone, was microinjected into C57/B16 host blastocysts. Chi-meric males were bred to NIH Swiss black females. Agoutioffspring were screened for germ-line transmission of thetargeted mutation by Southern blot analysis of SspI endonu-clease digested tail DNA using the 1.2-kb HindIII fragment asprobe (see Fig. 1). Fl heterozygous mice were bred to establisha colony of SP-A(+/+), SP-A(+/-), and SP-A(-/-) mice.

Immunohistochemistry, Western Blotting, and NorthernBlotting. Lungs from adult SP-A(+/-), (-/-), and (+/+)mice were immunostained for SP-A, proSP-B, and proSP-Cusing guinea pig anti-rat polyclonal SP-A and rabbit polyclonalproSP-B and proSP-C antiserum as described (15-17). Lungs

Abbreviations: ES cells, embryonic stem cells; Sat PC, saturatedphhosphatidylcholine.tTo whom reprint requests should be addressed at: Children's HospitalMedical Center, Division of Pulmonary Biology, The Children'sHospital Research Foundation, 3333 Burnet Avenue, Cincinnati, OH45229-3039.§Present address: Department of Environmental Medicine, Universityof Rochester, Rochester, NY 14642.

9594

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement" inaccordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Mar

ch 2

1, 2

020

Proc. Natl. Acad. Sci. USA 93 (1996) 9595

were prepared by inflation fixation and embedded in paraffinbefore sectioning. For Western blot analysis of lung homog-enates, mice were euthanized under protocols approved by theInstitutional Animal Use and Care Committee. Lungs werehomogenized in 10 mM Tris HCl (pH 7.4), 0.25 M sucrose, 2mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 10 ,tMleupeptin, and 10 ,uM pepstatin A and centrifuged at 250 x gfor 10 min at 2°C. Supernatants were centrifuged at 120,000 xg for 18 hr and the pellet resuspended in homogenizing buffer(minus sucrose) and subjected to SDS/PAGE on 10-27%gradient gels and transblotted to polyvinylidene difluoride(Bio-Rad) membranes, then blocked with 10mM Tris HCl (pH7.4), 0.15 M NaCl, 0.1% Tween 20 (TBS-T) containing 5%bovine serum albumin. Blots were incubated with the anti-surfactant protein antisera described above or with anti SP-Dkindly provided by E. Crouch (St. Louis, MO). Blots wererinsed and an appropriate horseradish-peroxidase-conjugatedsecondary antibody (Calbiochem) was added for 4 hr. Blotswere rinsed and developed using ECL detection reagents(Amersham). Immunoreactive bands were identified by ex-posing blots to XAR film (Kodak). Where indicated, thepolyvinylidene difluoride membrane was stripped using 62.5mM Tris-HCl (pH 6.8), 2% SDS, and 0.1 M 2-mercaptoetha-nol, and reanalyzed using another primary antiserum. RNAisolation, Northern blot analyses, and Si nuclease analyseswere performed as described (14, 18, 19).

Pressure-Volume Curves. Lung compliance was measuredin SP-A(+/+), (+/-), and (-/-) mice. Mice were injectedi.p. with sodium pentobarbital (200 mg/kg) and placed in achamber containing 100% 02 to completely deflate the lungs.A catheter (24 g) was inserted into the trachea and connectedto a silicon pressure sensor (X-ducer; Motorola, Phoenix, AZ).The chest wall and diaphragm were opened, and the lungs wereinflated with air in small increments (0.1 ml every 20-30 s) toa maximal pressure of 25 cm H20. Lungs were deflated instepwise fashion. Airway opening pressure and lung volumewere recorded at each inflation and deflation increment. Allmeasurements were completed within 15 min. Pressure-volume curves from each animal were used to assess lungcompliance, defined as the slope of the linear region of thedeflation curve where pressure ranged from -10 to 10 cmH20. Specific lung compliance was calculated as lung compli-ance divided by lung volume at 25 cm H20. The hysteresis ratiowas determined as described (20), using software kindly pro-vided by Gary S. Huvard (Huvard Research and Consulting,Chesterfield, VA). Differences among animals were evaluatedby comparing lung compliances, specific lung compliances,hysteresis ratios, and lung volumes at 25 cm H20 by one-wayanalysis of variance.

Electron Microscopy. Lungs from each of two SP-A-deficient (-/-) and two wild-type (+/+) mice were inflationfixed and prepared for electron microscopy as described (19).Ultrathin sections from 4-6 different blocks, representingboth proximal and distal portions of the lung, were examinedfor secreted forms of surfactant phospholipid and type II cellmorphology. Five randomly selected contiguous fields (equiv-alent to 19,845 ,um2) from each section were then scored at amagnification of 80,000 for the presence of tubular myelin. Atotal of 40 fields from the wild-type mice (158,760 ,um2) and70 fields from the SP-A-deficient mice (277,830 gm2) wereexamined to determine the relative abundance of tubularmyelin in these samples.

Alveolar Lavage and Surfactant Characterization. Micewere injected i.p. with pentobarbital to achieve deep anesthe-sia, and the distal aorta was cut to exsanguinate each animal.For alveolar lavage, the chest of the animal was opened, a 20 gblunt needle was tied into the proximal trachea and five 1 mlaliquots of 0.9% NaCl were flushed into the lungs andwithdrawn by syringe three times for each aliquot. The recov-ered lavage was pooled and the volume measured. The amount

of saturated phosphatidylcholine (Sat PC) recovered by alve-olar lavage was measured by extracting an aliquot of eachalveolar lavage with chloroform/methanol (2:1), followed bytreatment of the lipid extract with OS04 in carbon tetrachlo-ride and alumina column chromatography according to Masonet al. (21). Phosphorus in Sat PC was measured by the Bartlettassay (22). Lung tissue after alveolar lavage was homogenizedin saline and an aliquot was extracted and analyzed for Sat PCcontent. For measurements of phospholipid composition,chloroform/methanol extracts of alveolar lavages of individualmice were used for two-dimensional thin-layer chromatogra-phy (23). The spots were visualized with iodine vapor, scraped,and assayed for phosphorus content.

Surface Tension Measurements. Alveolar lavages were col-lected as above from 7- to 8-week-old mice by a series of fivesequential lavages. Because surface tension measurementswere made in the presence and absence of Ca2 , separatepooled lavages were recovered using 0.9% NaCl or a calciumcontaining buffer (0.01 M Tris HCl/0.0025 M CaCl2/0.001 MMgSO4/0.0001 M EDTA/0.15 M NaCl, pH 7.0). NaCl (0.9%)or the Ca2+ containing buffer was used for all subsequentprocessing and surface tension measurements. Surfactant wasisolated by separating large aggregate and small aggregatefractions by centrifugation as described (24). The alveolarlavages were centrifuged at 40,000 x g for 15 min and the pelletwas suspended in 0.9% NaCl or Ca2+ buffer and centrifugedagain at 40,000 x g for 15 min. The surfactant pellet wassuspended in a small amount of 0.9% NaCl or Ca2+ buffer. SatPC in aliquots of large and small aggregate surfactant weremeasured and the percentage of large aggregate surfactant inalveolar lavage was calculated. The minimum surface tensionwas measured with a Wilhelmy balance with a platinum

A

H

5.6

4.1

S N N

H pGKneoppa X

H

BH

probe 2

+1+ +1 - -1l-

FIG. 1. SP-A locus targeting construct. (A) The top line (-4.8 kb)depicts the mouse SP-A locus (14) with exons represented as boxes.Only the translated portion of exon 6 is depicted. pGKneoBPA wasinserted into the 3' terminus of exon 3 and 5' terminus of exon 4,deleting those portions of the exons and intron 3. Pertinent restrictionendonuclease sites and the position of probe 2 are depicted. Probe 1(not shown) was used with EcoRI digestions, probe 2 (shown) withSspI, and probe 3 was the 1.5-Kb BamHI fragment. The targetingconstruct was the modified Hindlll fragment marked with asterisks. H,HindIII; S, SspI; B, BamHI; N, NcoI. (B) Southern blot of genomicDNA probed with probe 2 following SspI digestion. The 4.1-kb Ssplband is the normal allele and the 5.6-kb SspI band is the targeted allele.Wild type (+/+), heterozygous (+/-), and homozygous (-/-) aredepicted below the lanes.

B

Cell Biology: Korfhagen et aL

W.

..........

Dow

nloa

ded

by g

uest

on

Mar

ch 2

1, 2

020

9596 Cell Biology: Korfhagen et al.

+1+ +1+ /I /I3.0kb

SP-A 1.7

0.9*

neo l.Okb _.1

FIG. 2. SP-A mRNA in wild-type and SP-A(-/-) mice. Total lungRNA was probed with 32P-labeled SP-A or neomycin probes. Geno-type is depicted above the lanes. Size (in kb) of the expected mRNAbands are noted on the left. The data are representative of Northernblot analysis of lung RNA from five separate mice of both genotypes.

dipping plate using 3 min area cycling from 64 cm2 to 12.8cm2 at a temperature of 37°C (25). Surfactant was diluted in35 ml 0.9% NaCl or Ca2+ buffer to a concentration of 0.005,umol Sat PC/ml and, after each surface tension measure-ment, increments of 0.005 ,umol Sat PC were added tomaximum concentration of 0.025 ,umol Sat PC/ml. Thesurfactant suspension was mixed after each surfactant ad-dition, and surface compression was started 20 sec later. Thesurface tension-area loops overlapped by the third or fourthcompression cycle and minimum surface tension on fourthcycle is reported.

Ab

anti-SP-A

anti-SP-B

rat mouse genotypeSP-A -/- +/+ +1-

U. 1.

anti-SP-D

FIG. 3. Immunoblot of SP-A; proSP-B and SP-D in lung homog-enates. Lung homogenates were prepared from SP-A(+ / +), SP-A(+ /-), and SP-A(-/-) mice, subjected to SDS/PAGE, and blotted asdescribed. Antibody (Ab) is noted on the left; genotype is markedabove the lanes. SP-A (34 kDa), proSP-B (-42 kDa), and SP-D (-43kDa) are noted. Data are representative of two to five separateexperiments. Rat SP-A was included in the gel as control.

RESULTSSouthern blot and partial restriction and sequence analysis ofthe 129J murine genomic SP-A locus demonstrated that SP-Awas encoded by a single gene that was indistinguishable fromthe murine SP-A gene described in strain DBA/2J (14). The129J genomic clone was used to prepare a targeting constructas depicted in Fig. 1. Insertion of pGKneoBPA deleted aportion of exon 3, intron 3, and a portion of exon 4 (fromposition 795-1049). Although the resulting allele maintainstranscriptional start sites, the pGKneoBPA insert provides astrong competing start site and polyadenylylation signal andremoves the splice signals of intron 3. A targeted ES cell clone

1!

j-._

I=t

1%%.

.s

14,~~~~~~~~~~~~,S 'Xl

100

t s -;-T kr . i*...r

Lk

4.. "

*.I

':

D t,NAe A X

-. 4

..

FIG. 4. Immunohistochemistry of surfactant proteins in SP-A(-/-) mice. Lung sections were stained for SP-A (A and C) and SP-B (B andD) from SP-A(-/-) (A and B) or SP-A(+/+) (C and D) as described. Histology and immunostaining of SP-A(+/-) mice were similar to thosein wild-type mice (data not shown). (Bar = 100 Am.)

f

pI

.

I .

.0

SI; N,

4.

Proc. Natl. Acad. Sci. USA 93 (1996)

LA 11

0 4

J"O! 4.J. OWN.Aw. t A

-,7 O"

*-...VWAll f -1

dp

310.

.o.,f.i4

'o fkil.,4 !f,:.:;",dlllllllik.L,wqm-jk..

.0

i"ir

.i:

Dow

nloa

ded

by g

uest

on

Mar

ch 2

1, 2

020

Proc. Natl. Acad. Sci. USA 93 (1996) 9597

was microinjected into mouse blastocysts that were transferredto surrogate dams. Chimeric males were bred to NIH Swissblack females. Heterozygous Fl offspring were bred to pro-duce SP-A (+/+), (+/-), and (-/-) mice. The distributionof genotype among the initial 246 offspring was 67 (+/+), 128(+/-), 51 (-/-) for a ratio of 0.54/1.04/0.42, approximatingthat predicted by normal Mendelian inheritance. Heterozy-gous and homozygous SP-A gene targeted mice were notdistinguishable from normal littermates in growth and repro-duction. A pair of SP-A(-/-) mice have been maintained inmicroisolator containment for more than 8 months withoutevidence of ill health.SP-A mRNA was not detectable in the lungs of the homozy-

gous gene targeted animals by Northern blot analysis (Fig. 2).Using S1 nuclease analysis, the abundance of SP-AmRNA wasreduced by approximately half in the heterozygous SP-A genetargeted animals (not shown). SP-A protein was not detectedin lung homogenates from homozygous mice (-/-; see Fig. 3).There was a reduction of SP-A protein in the heterozygousSP-A gene targeted animals compared with control litter-mates. SP-B and SP-D protein were not altered in SP-A(+/ -)or SP-A(-/-) mice (Fig. 3). ProSP-C protein was also notaltered (data not shown).

Histochemical analysis of the lungs from the SP-A(-/-)mice revealed no abnormalities of lung structure and noevidence of inflammation. Immunohistochemical stainingdemonstrated the lack of SP-A in alveolar type II cells ofSP-A(-/-) mice (Fig. 4A). The distribution of type II cellsand abundance of surfactant protein proSP-B was also un-changed in SP-A(-/-) mice (Fig. 4B). The distribution ofproSP-C was unchanged (not shown).

Composition of phosphatidylcholine, phosphatidylglycerol,phosphatidylinositol, phosphatidlyethanolamine, and sphingo-myelin was not altered in alveolar lavages or lung tissue ofSP-A(-/-) mice (Fig. 5). Lung volumes and compliance ofthe SP-A(-/-) mice were similar to wild-type SP-A(+/+)littermates (Table 1). The percentages of surfactant isolated aslarge aggregates in alveolar lavages were similar for bothgroups and were 34.0 ± 2.9% of the total Sat PC (n = 10 poolsof seven animals each) for SP-A(+/+) mice and 34.4 ± 2.7%(n = seven pools of seven animals each) for SP-A(-/-) mice.Independent of isolation or measurement of minimum surfacetension in the presence or absence of Ca2 , surfactant from

90*

60-

30-

0O

40-

20-

n

A

PC

GENOTYPE

+/++/-

- -mm

PG Pi PEPS

Table 1. Lung volumes and compliance

Genotype Lung volume Compliance* Hysteresis ratio

+/+ (n = 7) 28 ± 3 0.091 ± 0.008 0.383 ± 0.031+1- (n = 3) 26 ± 5 0.058 ± 0.008 0.439 ± 0.024-/- (n = 9) 26 + 3 0.074 ± 0.008 0.384 ± 0.024

Lung volume is expressed as ,ud contained in the lung at 25 cm H20divided by body weight. All values were compared by one way analysisof variance and are expressed as mean ± SEM.*Compliance was determined by calculating the slope of the deflationpressure-volume curve between -10 and + 10 cm ofwater divided bylung volume.

SP-A(+/+) and SP-A(-/-) mice had low minimum surfacetensions at a bulk phase concentration of 0.025 ,umol SatPC/ml (Fig. 6). Surface tensions for surfactant from SP-A(-/-) mice tended to be higher at lower concentrations inthe presence of Ca2+ and were significantly higher in theabsence of Ca2+ than for surfactant from SP-A(+/+) mice.To determine if the morphology of type II cells or tubular

myelin was altered, lungs were prepared for electron micros-copy. The morphology of type II epithelial cells in SP-A(-/-)was identical to that of wild-type animals (Fig. 7A). Whiletubular myelin structures were readily detected in wild-typelittermates, tubular myelin figures were markedly deficient inthe SP-A(-/-) mice. Of 70 fields examined from two (-/-)mice, only two small tubular myelin figures were detected; oneis shown in Fig. 7B. In contrast, 40 fields examined from two(+/+) mice revealed 15 large tubular myelin figures and onerepresentative figure is shown in Fig. 7C. The preponderantform of alveolar lipids in SP-A(-/-) mice were lamellatedand membranous lipid forms lacking the dense, highly geo-metric forms characteristic of tubular myelin Fig. 7D).

DISCUSSIONSP-A deficiency in homozygous SP-A gene targeted mice ledto production of surfactant with reduced surface tension

40

30

z0i75zww- E0 0

La.. cix4)

Sph Lyso-PC

B -I-

O~~~~

ALVEOLARWASH

LUNGTISSUE

TOTALLUNG

FIG. 5. Phospholipid composition. (A) Phospholipid compositionwas measured in alveolar lavage and lung homogenates from SP-A( + /+), SP-A(+/-), and SP-A(-/-) mice. Values are mean ± SEM. (B)Sat PC was determined in the alveolar washed lung tissue (mean ±

SEM; n = 4-6 determinations).

20

10-

2

A*

GENOTYPE0 +/+

T 0 -/-uI * *

0 0

i - i i 1T __T __III

u .I . .I

Bto - T

I T0n*

0

vz -_ .w

0.000 0.005 0.010 0.015 0.020 0.025

SAT PC (smol/ml)FIG. 6. Minimum surface tensions. Surfactants recovered from 10

mice for each of three pooled samples isolated in the presence orabsence of Ca2+ from SP-A(+ / +) and SP-A(- / -) mice were used forsurface tension measurements. Measurements were made in theabsence (A) or presence (B) of 2.5 mM Ca2+. Values are for threemeasurements at each concentration (mean SEM). *, P < 0.05relative to SP-A(+/+) mice.

0I0-

I

0j

0

0_s

E

0

0~

I-1CM

Cell Biology: Korfhagen et aL

-

1

Dow

nloa

ded

by g

uest

on

Mar

ch 2

1, 2

020

9598 Cell Biology: Korfhagen et al.

f(

1 urm

.t~ ~ ~ ~+k-; ' i.. /+C

FIG. 7. Ultrastructural analysis of type II alveolar cells and alveolar surfactant in SP-A-deficient mice. Lung tissue was obtained from adult miceand prepared for electron microscopy as described. (A) Intact, mature lamellar bodies (arrows) were readily detected in the apical cytoplasm oftype II alveolar cells from SP-A-deficient mice. Surfactant secretion was also observed at the apical cell surface (arrowhead). (Bar = 1.0 ,im.) (B)One of two tubular myelin figures detected in SP-A(-/-) mice. (C) One of 15 tubular myelin figures from SP-A(+/+) mice. (D) A representativelung section from SP-A(-/-) mice showing membranous forms of phospholipid that lacked the more highly ordered, geometric organizationcharacteristic of tubular myelin. (B-D were printed at the same magnification; bar in B is 1.0 um.)

lowering properties at low phospholipid concentrations with-out altering levels of intracellular/extracellular surfactantphospholipids. The abundance of tubular myelin figures in vivowas reduced but the abundance of surfactant proteins B, C,and D were unaltered. SP-A-deficient mice have no apparentbreeding or survival abnormalities when housed in well-controlled, pathogen-free cages in the vivarium. These findingsdemonstrate a phospholipid concentration dependence onsurface tension reducing properties in the absence of SP-A andthe importance of SP-A in the formation of tubular myelin, thecharacteristic extracellular form of surfactant. The findings ofsmall tubular myelin figures in the (- /-) mice may suggestthat figures are transiently formed but not stabilized in theabsence of SP-A. Alternatively, there may be very low levels ofSP-A, below detectability, that are sufficient to lead to for-mation of markedly reduced numbers of tubular myelin fig-ures.The disruption of the SP-A locus led to a lack of detectable

SP-A mRNA and protein in the SP-A(- /-) mice and was notassociated with changes in surfactant proteins SP-B, proSP-C,or SP-D. These findings support previous data suggesting thatsurfactant protein genes are independently regulated. As in theSP-B null mutant mice (19), wherein SP-A and -C mRNAswere unaltered, the regulation of surfactant protein genes doesnot appear to be under compensatory regulatory signalsrelated to deficiency of SP-A or SP-B.

Consistent with the finding that the concentration of SP-Band -C in lavage from the SP-A( - / -) mice were unaltered, the

distribution and abundance of immunostaining for surfactantproteins proSP-B and proSP-C were also not different fromcontrols. Immunohistologic analysis demonstrated the com-plete absence of SP-A in the alveoli of SP-A(-/-) and areduced amount in SP-A(-/+) mice, demonstrating that thelevel of SP-A does not alter the steady-state levels of SP-AmRNAs and protein. In spite of the reduction of SP-A inSP-A(+/-) and the absence of SP-A in SP-A(-/-) mice, thegene targeted animals continue to survive (at least 8 months)with no apparent illness, in viral free housing with controlledhumidity and temperature and access to unlimited food andwater.SP-A has been postulated to play an important role in the

regulation of surfactant phospholipid levels by inhibiting phos-pholipid secretion of type II cells and enhancing the reuptakeof surfactant phospholipids in vitro (4-6). However, in thepresent study, there were no detectable alterations in intra-cellular or extracellular pool sizes or composition of surfactantphospholipids from SP-A(-/-) mice suggesting that SP-Adoes not determine the abundance of the major phospholipidclasses of surfactant in vivo. Alternatively, counterregulatorymeasures, independent of surfactant proteins or pool sizes,may contribute to the maintenance of steady-state phospho-lipid concentrations in the intracellular and extracellularspaces in the SP-A(-/-) mice.At the highest bulk phase concentration tested on the

Wilhelmy surface balance, surfactant from the SP-A(-/-)mice yielded the same low surface tension as surfactant from

Proc. Natl. Acad. Sci. USA 93 (1996)

Dow

nloa

ded

by g

uest

on

Mar

ch 2

1, 2

020

Proc. Natl. Acad. Sci. USA 93 (1996) 9599

control mice. At lower surfactant concentrations, minimumsurface tensions are higher for surfactant from the SP-A(-/-) mice. These results are consistent with the lack ofapparent abnormalities in lung compliance in the SP-A(-/ -)mice because surfactant concentration is very high in thecontinuous thin film of fluid that is the alveolar lining layer(26). By extrapolation from estimates for the human, the adultmouse lung would have an alveolar fluid volume of 40 ,ul anda surface area of about 0.1 m2 (27, 28). The concentration ofSat PC in this alveolar fluid would be on the order of 10,umol/ml, a value more than three orders of magnitude higherthan tested on the surface balance. Phospholipids in thepresence of SP-B and SP-C without SP-A were previouslyshown to spread normally (29). The lack of altered surfactantfunction in the SP-A(-/-) mice is further supported by theobserved normal lung compliance (Table 1).

Surfactant homeostasis is mediated by complex synthetic,secretory, and catabolic processes that include recycling andcatabolism of surfactant phospholipids by respiratory epithe-lial cells and catabolism by alveolar macrophages. The findingsthat steady-state phospholipid concentrations were unalteredin the SP-A-deficient mice are consistent with previous find-ings demonstrating the lack of effect of SP-A on clearance ofsurfactant mixtures in vivo. In the neonatal rabbit, the clear-ance of surfactant lipids in the airspace was identical in thepresence or absence of exogenous SP-A (30). In contrast, atnonphysiologic concentrations of phospholipids, there is adifference in surface tension lowering properties of isolatedsurfactant. These data suggest that in conditions where phos-pholipid concentrations were altered in vivo, SP-A could playa critical role in maintenance of surface tension reducingproperties. Such conditions may occur with infections, toxic, orimmunologic injury to the lung epithelium. This concept issupported by in vitro evidence that SP-A protects surfactantfrom protein inactivation (9) and binds and enhances uptakeof microbial pathogens (10-13). In future studies it will beimportant to determine pulmonary function in SP-A-deficientmice following infectious or toxic injury to determine the fullphysiologic functions of SP-A.

We thank Dr. E. Crouch for the gift of an anti-SP-D antibody andDrs. John Duffy and Tom Doetschman for assistance in generatinggene targeted chimeric mice. This work was supported by GrantHL28623 from the National Institutes of Health (T.R.K. and J.A.W.),by the Cystic Fibrosis Foundation, and by the Program of ExcellenceMolecular Biology (Grants HL41496 and HD11932).

1. Moore, K. J., D'Amore-Bruno, M. A., Korfhagen, T. R., Glasser,S. W., Whitsett, J. A., Jenkins, N. A. & Copeland, N. G. (1992)Genomics 12, 388-393.

2. Hawgood, S. & Clements, J. A. (1990) J. Clin. Invest. 86, 1-6.3. White, R. T., Damm, D., Miller, J., Spratt, K., Schilling, J.,

Hawgood, S., Benson, B. & Cordell, B. (1985) Nature (London)317, 361-363.

4. Wright, J. R., Wager, R. E., Hawgood, S., Dobbs, L. & Clements,J. A. (1987) J. Biol. Chem. 262, 2888-2894.

5. Dobbs, L. G., Wright, J. R., Hawgood, S., Gonzales, R., Ven-strom, K. & Nellenbogen, J. (1987) Proc. Natl. Acad. Sci. USA 84,1010-1014.

6. Rice, W. R., Ross, G. F., Singleton, F. M., Dingle, S. & Whitsett,J. A. (1987) J. Appl. Physiol.. 63, 692-698.

7. Suzuki, Y., Fujita, Y. & Kogishi, K. (1989) Am. Rev. Respir. Dis.140, 75-81.

8. Hawgood, S., Benson, B. J. & Hamilton, R. L. J. (1985) Biochem-istry 24, 184-190.

9. Cockshutt, A. M., Weitz, J. & Possmayer, F. (1990) Biochemistry29, 8424-8429.

10. Gaynor, C., McCormack, F. X., Voelker, D. R. & Schlesinger L.(1995) J. Immunology 155, 5343-5351.

11. van Iwaarden, J. F., Pikaar, J. C., Storm, J., Brouwer, E., Ver-hoef, J., Oosting, R. S., Van Golde, L. M. G. & Van Strijp,J. A. G. (1994) Biochem. J. 303, 407-411.

12. Tenner, A. J., Robinson, S. L., Borchelt, J. & Wright, J. R. (1989)J. Biol. Chem. 264i 13923-13928.

13. Pison, U., Max, M., Neuendank, A., Weibbach, S. & Pietchmann,S. (1994) J. Clin. Invest. 24, 586-599.

14. Korfhagen, T. R., Bruno, M. D., Glasser, S. W., Ciraolo, P. J.,Whitsett, J. A., Lattier, D. L., Wikenheiser, K. A. & Clark, J. C.(1992) Am. J. Physiol. 263, L546-L554.

15. Slavkin, H. C, Johnson, R., Oliver, P., Bringas, P., Don-Wheeler,G., Mayo, M. & Whitsett, J. A. (1989) Differentiation 41,223-236.

16. Khoor, A., Stahlman, M. T., Gray, M. E. & Whitsett, J. A. (1994)J. Histochem. Cytochem. 42, 1187-1199.

17. Vorbroker, D. K., Profitt, S. A., Nogee, L. M. & Whitsett, J. A.(1995) Am. J. Physiol. 268, L647-L656.

18. Dranoff, G., Crawford, A. D., Sadelain, M., Ream, B., Rashid, A.,Bronson, R. T., Dickersin, G. R., Bachurski, C. J., Mark, E. L.,Whitsett, J. A. & Mulligan, R. C. (1994) Science 264, 713-716.

19. Clark, J. C., Wert, S. E., Bachurski, C. J., Stahlman, M. T., Stripp,B. R., Weaver, T. E. & Whitsett, J. A. (1995) Proc. Natl. Acad.Sci. USA 92, 7794-7798.

20. Nardell, E. A. & Brody, J. S. (1982) J. Appl. Physiol. 53, 140-148.21. Mason, R. J., Nellenbogen, J. & Clements, J. A. (1976) J. Lipid

Res. 17, 281-284.22. Bartlett, G. R. (1959) J. Biol. Chem. 234, 466-468.23. Jobe, A. H., Kirkpatrick, E. & Gluck, L. (1978)J. Biol. Chem. 253,

3810-3816.24. Ikegami, M., Ueda, T., Hull, W., Whitsett, J. A., Mulligan, R. C.,

Dranoff, G. & Jobe, A. H. (1996)Am. J. Physiol. 270, L650-L658.25. Ikegami, M., Jobe, A. & Glatz, T. (1981) J. Appl. Physiol. 51,

306-312.26. Bastacky, J., Lee, C. Y. C., Goerke, J., Koushafar, H., Yager, D.,

Kenaga, L., Speed, T. P., Chen, Y. & Clements, J. A. (1995)J. Appl. Physiol. 79, 1615-1628.

27. Rannard, S. I., Basset, G., Lecossier, D., O'Donnell, K. M.,Pinkston, P., Martin, P. G. & Crystal, R. G. (1986) J. Appl.Physiol. 60, 532-538.

28. Weibel, E. R. (1987) Annu. Rev. Physiol. 49, 147-159.29. Ross, G. F., Notter, R. H., Meuth, J. & Whitsett, J. A. (1986)

J. Biol. Chem. 261, 14283-14291.30. Ikegami, M., Lewis, J. F., Tabor, B., Rider, E. D. & Jobe, A. H.

(1992) Am. J. Physiol. 262, L765-L772.

Cell Biology: Korfhagen et al.

Dow

nloa

ded

by g

uest

on

Mar

ch 2

1, 2

020